Chief Investigators: Dr Philippa Nicolson (University of Oxford) and Prof Melanie Holden (Keele University)
Background
In collaboration with the OA Trial Bank, we recently undertook Individual Participant Data (IPD) meta-analyses using data from randomised controlled trials (RCTs) comparing therapeutic exercise to non-exercise controls among people with knee and/or hip OA (the STEER OA study (Holden et al 2023). Our findings suggest that targeting exercise to those with higher OA-related pain and disability might be of merit. However, the heterogeneity of exercise interventions included in our analyses means there is considerable uncertainty about the relative effectiveness of different characteristics of exercise interventions. The optimal type and mode of delivery was not explored in the STEER OA study. This makes it difficult to know how to optimise the provision of therapeutic exercise for people with knee and/or hip OA.
Aim
This study aims to summarise, compare and rank the effectiveness of different types and modes of delivery of therapeutic exercise on pain and physical function at 3, 6 and 12 months for people with knee and/or hip OA in an IPD network meta-analysis (NMA), in terms of: (a) the overall effect (across all individuals); and b) the effect in specific subgroups that cause heterogeneity in effects (effect moderators) if inconsistency of evidence is identified in analysis.
Methods
Systematic review update and IPD meta-analyses. Our previous systematic review (Holden et al 2023) will be updated to identify RCTs that compare the effects of therapeutic exercise for people with knee and hip OA on pain and physical function, to other forms of exercise or non-exercise controls. Lead authors of eligible trials will be invited to share IPD. Trial- and participant-level characteristics (for baseline variables and outcomes) of included studies will be summarised. Meta-analyses will use a two-stage approach: in the first stage, each RCT will be analysed separately to produce RCT-specific estimates. In the second stage, a multivariate NMA model will be fitted for each outcome separately, allowing for exercise type and mode of delivery for each time-point, whilst accounting for the correlation across time-points in a multivariate model. The models will include random-effects to allow for unexplained between-trial heterogeneity. All analyses will be on an intention-to-treat principle and all summary meta-analysis estimates will be reported as standardised mean differences with 95% confidence intervals.
Status
Ongoing.